Re: [PATCH] [gdb/testsuite] Fix another fail and tcl error in gdb.dap/sources.exp

[Tom de Vries <tdevries@suse.de>]

On 2/13/24 19:08, Tom Tromey wrote:
>>>>>> "Tom" == Tom de Vries <tdevries@suse.de> writes:
> 
> Tom> These are the same type of tcl error and FAIL I just fixed for a later
> Tom> request in the same test-case.
> 
> Tom> Fix this by:
> Tom> - moving the wait-for-stop to before the loadedSources request to fix the
> Tom>   FAIL, and
> Tom> - checking for $obj == "" to fix the tcl error.
> 
> This is fine as long as these tests are actually being run in the end.
> What I mean is, it would be bad to cover up a real failure of some kind.
> 

Hi Tom,

thanks for the review.

Agreed, that would be bad, but I don't think this is the case.

The only thing that's happening here is that the test-case is cut short 
as soon as one test FAILs. There's always a FAIL to signal that 
something went wrong.

The dap_check_request_and_response proc itself produces a FAIL, which is 
why it isn't obvious from looking at the test-case that a FAIL is produced.

[ I'm on the fence about this.  Having the proc itself produce the fail 
guarantees a FAIL without effort from the call, but make you wonder when 
looking at the call whether something is missing.  Conversely, having to 
check the result of the call and produce a FAIL means it can be forgotten. ]

I know there's some high level goal to produce the same amount of 
PASS/FAIL, but IMO that's to be applied if that's easy, if it makes the 
test-case convoluted then it's not worth the trouble.

Committed.

Thanks,
- Tom

> Approved-By: Tom Tromey <tom@tromey.com>
> 
> Tom